Metagenomics – methods for sequencing environmental samples with an intermediary cultural step – have revolutionized our understanding of how microbial life shapes and is shaped by their environments, with important implications for human health and environmental monitoring. I work to develop methods for analyzing this data. My most recent paper on the lineage model (figure at left), together with Daniel Falush and Xavier Didelot, shows how to infer the community structure and clonal history for groups of closely-related species using shotgun metagenomic data. I am currently working to extend methods for amplicon metagenomic and metatranscriptomic approaches.
Mixtures of P. falciparum
Plasmodium falciparum, the causative agent of most severe malaria, shows incredible levels of mixing within clinical isolates. A single infected individual may posses a significant fraction of all of the regional variation present. The level of mixture is implicated in clinical progress, the rate of outcrossing in the P. falciparum, and drug resistance emergence. I am particularly interested in how to measure the level of mixture within clinical samples that have been sequenced using NGS techniques. However, measuring how mixed a sample appears is deeply tied to the model of how one thinks mixture occurs. Together with Lucas Amenga-Etego, I am working to extend the initial, frequentist metrics to model-based approaches that compare different biologically-realistic models to give clinicians a more rich understanding of how mixture operates.
Like many applied statisticians, I’ve dabbled with many data sets, largely from student projects. They’ve included work on RFID data from social wasps, Twitter temporal patterns, recovering social networks from card-swipe data, and many ecological and genetic data sets. If you’re interested in any of these, please get in touch!